3. Data, Science & AI

Guides effective design of AI agents through prompt engineering, tool selection, and permission scoping for secure and efficient operation.

Manages TimescaleDB hypertables, continuous aggregates, and schema migrations for time-series data workloads.

Provides generic patterns for database queries, migrations, and performance optimization applicable across any database and ORM.

Guides integration of external economic and cryptocurrency data sources (FRED, CoinGecko) into background workers for dashboard data expansion.

Designs optimized prompts for Google Gemini API to produce structured JSON outputs, enhancing AI interaction reliability for TalentFilter applications.

Manages AI agent context via Agentic File System (AFS) for discovery, mounts, scratchpad, and tool workflows.

Validates AI prompt files against reference guidelines to ensure consistency and adherence to defined standards.

Manages training, evaluation, and deployment of persona-based avatar models including voice, dataset preparation, and A/B testing for conversational AI agents.

AI navigation agent for The Legend of Zelda: Oracle of Secrets, using map data and dynamic state to route Link.

Explains Euler's modular oracle infrastructure and its integration with price feed sources including Chainlink, Uniswap TWAP, and Pyth.

Optimizes LLM conversation context by progressive disclosure and tiered loading to avoid token limits and improve efficiency in long interactions.

Guides creation of reusable, secure AI agent skills via a 10-step design loop with progressive disclosure and VOI-driven uncertainty management.

Manages AI model training, evaluation, and deployment (GGUF/MLX) for Zelda-themed variants, including dataset curation and registry updates.

Maps natural language queries to complex RAM inspections, enabling intuitive state queries for the Oracle of Secrets system.

Provides standardized execution rules for AI agents including TDD cycles, quality assurance, context management, and completion reporting across all development tasks.

Checks progress of Hive data tests and automatically updates documentation with estimated completion times.

Sets up Langfuse datasets with evaluation dimensions, score configurations, and judge prompts for LLM model assessment.

An AI agent that continuously enhances performance by learning from failures, user corrections, and capability gaps using memory ranking for improved decision-making.

Manages human annotations and scores for Langfuse traces, including workflow management, pending review tracking, and export capabilities.

Manages a small-scale LLM project for Dart/Flutter learning, handling data preparation, model training, and project administration.

Manages healthcare analytics indicators and metrics within the Engage Analytics dbt project, including metric definitions and questionnaire models.

Multi-agent orchestrator for Claude Code, managing agent lifecycle, workspace setup, crash recovery, and CLI operations for coordinated AI workflows.

Selects and delivers optimal AI prompts for user tasks, enabling precise language model interactions and effective AI utilization.

Displays real-time status of the QAVR AI system, including operational mode, interaction metrics, top memory retrievals by Q-value, and configuration details.